RAM Spike
Problem reported by J Lee - 11/12/2025 at 9:49 AM
Resolved
Hi 
Is anyone else noticing a RAM spike on Build 9434? I first noticed it on Build 9420. Rebooting once every 3 or 4 days keeps the issue in check, but if there is no reboot, the RAM continues higher and higher. 

SmarterMail 9434
Windows Std 2019

J. Sebastian Lee Service2Client LLC 6333 E Mockingbird Ste 147 Dallas, TX 75214 - 877.251.3273

Richard Laliberte Replied
I remember a while back, earlier this year maybe, some people mentioning the same type of issue. At that time one of the members created and shared an automated script that could be used. I can't find it right now, but hopefully someone still has it.

Obviously one of those things that got fixed then got un-fixed for some reason lol i was waiting for this weeks update (since it's supposed to be very large) to upgrade from 9406, but if this issues is popping up again guess i'll be waiting longer ...
J Lee Replied
Hi Richard 
Thank you for the reply. I'm still seeing this issue on 9448
J. Sebastian Lee Service2Client LLC 6333 E Mockingbird Ste 147 Dallas, TX 75214 - 877.251.3273
Gerardo Altman Replied
Are you running a VM or bare metal?
If VM whats the hypervisor ?

""G
Jason Replied
We have been seeing this issue since October 2024. Support informed us at the end of October 2025 that they were testing an in house server with the same issue. Unfortunately, we are still struggling with the same behavior. Several of our servers must be rebooted every 1–4 days. This is happening on Linux as well as Windows servers.
J Lee Replied
Bare Metal, Yes SM had this issue a few years ago. Ours if not rebooted will reach 90+% RAM usage and make the server unusable. After reboot it drops back down to 20% - 30%. It seems that new work is staying in RAM after completed.
J. Sebastian Lee Service2Client LLC 6333 E Mockingbird Ste 147 Dallas, TX 75214 - 877.251.3273
Gerardo Altman Replied
Interesting.

something to note when running in KVM there is a known issue with QEMU and De-frag/ reclaim consuming all the memory.

i cant say that this may also be an issue on physical hardware but you never know.

posted a link just in case its of any help.

J Lee Replied
This appears to be a memory cache clearing issue, or not clearing. 
J. Sebastian Lee Service2Client LLC 6333 E Mockingbird Ste 147 Dallas, TX 75214 - 877.251.3273
Gerardo Altman Replied
@j lee

Was/ is there a fix?
Andrew Barker Replied
Employee Post Marked As Resolution
Gerardo,

Build 9476 includes changes that address the gradual memory leak.
Andrew Barker Senior Software Developer SmarterTools Inc. www.smartertools.com
Jason Replied
We have build 9476 installed on three servers. Two of the servers seem stable now. One server still appears to have the same gradual memory leak. It has had to be restarted once.  We are going to monitor it, and if it happens again, we will reopen the ticket.
Jason Replied

I wanted to update this thread. We applied build 9476 to three servers. After the update to 9476, two of the servers were stable. The server that we originally opened a ticket with support last year continued to experience the RAM issue.

We compared the servers and found that a large number of JSON files on the problematic server contained BOM characters, which I would assume came from migrating this server from Windows in order to see if the mem leak continued with Linux. The servers that were operating correctly did not have BOM characters in the JSON files. We corrected all affected JSON files, but this did not resolve the problem.

The issue was ultimately resolved by stopping the Index service, removing the IndexV2 folders for every user on the affected server, and then restarting the Index service. After doing this, the RAM leak stopped and the server has remained stable.

Hopefully this information helps anyone who is still experiencing this issue.

J Lee Replied
Nice work 
J. Sebastian Lee Service2Client LLC 6333 E Mockingbird Ste 147 Dallas, TX 75214 - 877.251.3273
Gerardo Altman Replied
@Jason Great sleuthing :)

Thanks for sharing!

We are planning the switch to Linux in the new year - any tips?

""Cheers
G
terry fairbrother Replied
"We are planning the switch to Linux in the new year - any tips?"

I was trialling on Linux on a remote test server. I first started on Mint desktop and noticed errors in the cyren folders. After emailing Logan, he recommended Ubuntu LTS. I retried with 'Server' with a gui, but had linux based issues so went with Ubuntu desktop and has been stable. I have actually migrated our own exchange server over to SM, so we are, essentially live on Linux (Hyperv VM)

Everything seems to work. Only one thing has popped up which might be Linux specific (or bug) is the date format when impersonating a user is in US date format, but the user and system are set to UK date format but shows correctly on webmail. 

Impersonating...

webmail....


Just a minor difference!!!
Jason Replied

I spoke too soon. Five days plus and things were looking good, but I came in this morning to take a look and RAM is starting to climb again. This also seems to follow a pattern. Is there anything that SmarterMail runs from Sunday into Monday that is different than the rest of the week? For example, some sort of housekeeping, log rotation, etc.

I am going to run a DotMem snapshot tonight and re-open the ticket with SmarterTools. Another one of the three servers is also showing RAM creep, but it is not nearly as bad. I will continue to monitor that one to see if it stabilizes or keeps climbing.

Reply to Thread

Enter the verification text